A Novel Quantitative Proteomics Strategy To Study

interactions with more than one phosphorylation site involved. .... each sample and incubated overnight at 37 °C. iTRAQ Labeling. The resulting ... l...
1 downloads 0 Views 298KB Size
A Novel Quantitative Proteomics Strategy To Study Phosphorylation-Dependent Peptide-Protein Interactions Fei Zhou,† Jacob Galan,‡ Robert L. Geahlen,† and W. Andy Tao*,†,‡ Department of Medicinal Chemistry and Molecular Pharmacology, and Department of Biochemistry and Bindley Bioscience Center, Purdue University, West Lafayette, Indiana 47907 Received June 14, 2006

Phosphorylation-dependent protein-protein interactions provide the mechanism for a large number of intracellular signal transduction pathways. One of the goals of signal transduction research is to understand more precisely the nature of these phosphorylation-dependent interactions. Here, we report a novel strategy based on quantitative proteomics that allows for the rapid analysis of peptide-protein interactions with more than one phosphorylation site involved. The phosphorylation of two tyrosine residues, Y342 and Y346, within the linker B region of the protein-tyrosine kinase Syk is important for optimal signaling from the B cell receptor for antigen. We employed four amino-specific, isobaric reagents to differentially label proteins interacting in vitro with four Syk peptides containing none, one, or two phosphates on tyrosine residues Y342 and Y346, respectively. In total, 76 proteins were identified and quantified, 11 of which were dependent on the phosphorylation of individual tyrosine residues. One of the proteins, peroxiredoxin 1, preferably bound to phosphorylated Y346, which was further verified by Western blotting results. Thus, we demonstrate that the use of 4-fold multiplexing allows for relative protein measurements simultaneously for the identification of interacting proteins dependent on the phosphorylation of specific residues. Keywords: Quantitative proteomics • Mass spectrometry • Isotope labeling • Phosphorylation • Tyrosine kinases

Introduction Protein-protein interactions affect all biological processes in a cell. Proteins rarely function individually, and it has been proposed that all proteins in a given cell are connected through an extensive network, where noncovalent interactions are continuously associating and dissociating. Dynamic proteinprotein interactions are frequently regulated by post-translational modifications (PTMs), most notably phosphorylation. Eukaryotic cells widely use reversible phosphorylation to regulate protein-protein interactions to transmit and integrate signals received from their environment.1 Phosphorylation-dependent protein-protein interactions have been studied using yeast and bacterial two-hybrid systems in which phosphorylation influences protein-protein associations to activate transcriptional readouts.2,3 Affinity purification of tyrosine-phosphorylated proteins also has been widely used to probe signaling pathways involved in cellular responses to a particular stimulus.4 This method is typically based on the binding of phosphorylated signaling proteins to modular protein domains (e.g., Src homology 2 (SH2)). Mann and coworkers5 have developed a proteomic approach using immobilized phosphopeptides for the analysis of peptide-protein * Corresponding author. E-mail: [email protected]. † Department of Medicinal Chemistry and Molecular Pharmacology, Purdue University. ‡ Department of Biochemistry and Bindley Bioscience Center, Purdue University 10.1021/pr0602904 CCC: $37.00

 2007 American Chemical Society

interactions. They combined immunoprecipitation of purified tyrosine-phosphorylated proteins with this approach to identify phosphospecific binding partners involved in fibroblast growth factor signaling.5 The approach takes advantage of the concept that important protein-protein interactions in cell signaling are frequently mediated by short, unstructured sequences that specifically interact with peptide binding domains. In the latter two examples, mass spectrometry played an important role in the identification of signaling proteins and the location of sites of phosphorylation. Quantitative proteomic techniques can be employed to obtain quantitative information with regard to phosphorylation status and signal-dependent protein interactions in signal transduction cascades.6 Quantitative proteomics also has been applied to biomarker discovery and the identification of specific components of protein complex.7-9 The principle of this method is based on the enrichment factor observed between two related samples differentially labeled with stable isotopes.10,11 As indicated by the relative intensity ratio on the mass spectra, the proteins of interest can easily be distinguished from the background proteins in the differentially labeled samples. Stable isotope tags have been introduced into proteins/peptides through in vivo metabolic labeling or in vitro chemical or enzymatic reactions. Among various isotopic labeling techniques, isobaric tagging for relative and absolute quantitation (iTRAQ) is a novel proteomic approach that has been used to quantify differences in protein levels at the tandem mass Journal of Proteome Research 2007, 6, 133-140

133

Published on Web 11/17/2006

research articles spectrometry (MS/MS) stage.12 The iTRAQ reagents consist of a reporter group, a balance group, and an amino reactive group. Up to four different protein samples can be quantified simultaneously with the multiplexed set of isobaric reagents. Peptides of the same sequence labeled with iTRAQ reagents are identical in mass in single MS mode, but in MS/MS mode produce strong, diagnostic, low-mass signature ions (m/z 114-117) suitable for quantitation. Protein-protein interactions in cell signaling are often mediated by short sequences of amino acids containing a phosphorylated residue.13 However, within a short amino acid sequence, there may exist multiple phosphorylation sites, and each individual site can play a critical role required for a specific interaction. One example is the protein tyrosine kinase, Syk. Syk is a 72-kDa nonreceptor protein tyrosine kinase with an N-terminal, tandem pair of SH2 domains separated by a long linker (linker B region) from a C-terminal catalytic domain.14 Syk plays a critical role in immune cells by coupling immune recognition receptors with multiple signaling pathways.15-18 There are numerous sites of tyrosine phosphorylation on Syk including Y342 and Y346 within linker B, both of which are important for optimal signaling.19 Several studies have identified effector proteins with SH2 domains that bind to one phosphorylation site, the other, or both within this region.20-25 Therefore, there is considerable interest in determining how each phosphotyrosine residue influences the interactions between Syk and its association with other proteins and how this association regulates the propagation of signals through phosphotyrosine signaling cascades. However, quantitative analyses of such interactions are complicated by the possibility of multiple combinations of states of phosphorylation that can occur at these two closely spaced sites. To establish a strategy based on quantitative proteomics, we took advantage of iTRAQ reagents that are capable of analyzing four samples simultaneously so that we were able to examine interactions in parallel at these two residues at different phosphorylation states. We prepared four Syk-derived peptides with the two tyrosine residues, Y342 and Y346, differentially phosphorylated to serve as baits in affinity pull-down experiments. We effectively identified 76 proteins, of which 11 were dependent on the status of phosphorylation of individual tyrosines. One of the proteins was verified by Western blotting. We demonstrate the use of 4-fold multiplexing to enable relative protein measurements simultaneously and to distinguish interacting proteins dependent on the phosphorylation of specific residues.

Materials and Methods Phosphopeptide Affinity Isolations. The peptides DTEVYESPYADPEEIR (YY), DTEVpYESPYADPEEIR (pYY), DTEVYESPpYADPEEIR (YpY), and DTEVpYESPpYADPEEIR (pYpY) were purchased from SynPep (Dublin, CA). Peptides contained C-terminal amides and were covalently coupled to Affi-Gel-10 (Bio-Rad) through the N-terminal amine. DG75 B cells were cultured in RPMI 1640 media supplemented with 10% heatinactivated fetal calf serum, 1% chicken serum, 50 µM 2-mercaptoethanol, 1 mM sodium pyruvate, 100 IU/mL penicillin G, and 100 µg/mL streptomycin. Cells (2 × 107)(were lysed in 1 mL of buffer containing 20 mM Tris (pH 7.5), 150 mM NaCl, 5 mM EDTA, and 1% (v/v) Triton X-100. Lysates were precleared by incubation with 20 µL of the resin without peptide for 1 h at 4 °C. Unbound proteins were then adsorbed onto 20 µL of each peptide-resin at 4 °C for 2 h. Resins were washed three 134

Journal of Proteome Research • Vol. 6, No. 1, 2007

Zhou et al.

times with lysis buffer and two times with deionized water. Bound proteins were eluted in 100 µL of 5% acetic acid at room temperature for 30 min, partially dried in a speedvac, and reconstituted in 50 mM trimethylammonium bicarbonate (pH 8.5). One microgram of sequence-grade trypsin was added into each sample and incubated overnight at 37 °C. iTRAQ Labeling. The resulting peptide samples were dried in a Speedvac and resuspended in 30 µL of 100 mM trimethylammonium bicarbonate (pH 8.5). iTRAQ reagents 114, 115, 116, and 117 were resuspended in 70 µL of ethanol and added to the peptide samples resulting from affinity purification using the immobilized peptides YY, pYY, YpY, and pYpY, respectively. Samples were incubated for 1 h at room temperature, and then 10 mM glycine was added for 15 min to quench unreacted iTRAQ reagents. The samples were pooled and dried under vacuum for analysis by MS. Offline Sample Isolation and MS Data Acquisition. The modified peptides were separated offline by microcapillary liquid chromatography (Agilent 1100, Palo Alto, CA) and analyzed by a MALDI-TOF/TOF mass spectrometer (ABI 4700, Foster City, CA). Peptides on the MALDI plate were detected, isolated, and fragmented in a completely automated fashion. MALDI matrix [75% acetonitrile (AcN), 25% water, 0.1% trifluoroacetic acid (TFA), and 3 mg/mL R-cyano-4-hydroxycinnamic acid (CHCA)] was delivered by a syringe pump (Harvard Apparatus, So. Natick, MA) at 800 nL/min and mixed postcolumn with the LC effluent (600 nL min-1) via a Upchurch T connector. The mixtures were spotted on a 144-spot MALDI target by the Agilent micro fraction collector following a predefined 12 × 12 array every 20 s during 60 min gradient. All MS/MS data were searched against the human database using the SEQUEST algorithm.26 Static modifications were made on Lys residues and N-termini (+144.1). Peptide mass tolerance was set at 0.1 Da. Peptide and protein probability scores were used to filter the data.27,28 Quantitative information was obtained using the Libra program, which is one component of open-source proteomics software (Trans-peptide pipeline, TPP, developed by the Institute for Systems Biology, Seattle, WA; http://www.proteomecenter.org). Although many peptide signals appeared in adjacent spots, quantitative information was acquired only on the same spot where the sequence of the peptide was identified. This is sufficient to obtain accurate quantitative information, since peptides of same sequence labeled with different iTRAQ tags coelute on the LC column.12 SDS-PAGE Analysis and Western Blotting. Detergent lysates from DG75 B cells were prepared and adsorbed onto resins containing the four immobilized peptides as described above. Bound proteins were eluted in SDS sample buffer, separated by 10-12% SDS-PAGE gel, and detected by silver staining. For Western blot analysis, proteins were transferred onto polyvinylidene difluoride (PVDF) membranes, which were probed with an anti-peroxiredoxin I antibody (LabFrontier, Seoul, Korea).

Results Strategy To Study Peptide-Protein Interactions Involving More Than One Phosphorylation Site. To investigate the role of individual phosphorylation sites in peptide-protein interactions, we designed a quantitative proteomics strategy illustrated in Figure 1. Four peptides corresponding in sequence to Syk amino acid residues 338-353, DTEVYESPYADPEEIR, and containing either no phosphate group (YY), a single phosphate group at Y342 (pYY), a single phosphate group at Y346 (YpY),

Study of Phosphorylation-Dependent Peptide-Protein Interactions

research articles

Figure 1. Scheme illustrating the proteomic approach to peptide-protein interactions dependent on multiple phosphorylation events. See text for details.

or phosphate groups at both positions (pYpY) were covalently immobilized on the solid phase through the N-terminal amine and then incubated with equal amounts of detergent lysates of human DG75 B cells. The interacting proteins were eluted from each resin and digested with trypsin. Peptides resulting from the affinity purification with YY, pYY, YpY, and pYpY as baits were labeled with iTRAQ-114, -115, -116, and -117 reagents, respectively. After quenching the reactions with glycine, the labeled peptides were combined for offline microcapillary LC separation, followed by MALDI-TOF/TOF analysis. Because the peptide samples were labeled with isobaric tags, any given peptide labeled with each of the four tags had the same nominal mass. Therefore, on single MS mode, every peptide peak was the sum of all samples in the mixture, and there was no splitting of MS precursor signals and no increase in spectral complexity when the samples were analyzed together (Figure 1). Accurate quantitative information can be obtained in MS/ MS mode based on the relative abundance ratio of strong, diagnostic, low-mass MS/MS signature fragment ions (m/z 114-117) produced from iTRAQ-labeled peptides. Background or nonspecifically adsorbed proteins will be present in equal amounts in all four samples, and therefore, peptides from such proteins will result in equal abundance of ions m/z 114-117 in the MS/MS spectra. If a protein binds more specifically to one individual phosphorylated peptide, then MS/MS of peptides from the protein lead to a ratio of m/z 114-117 that is

related to the specificity of the binding to one form over the others. The simultaneous analysis of four samples allows for the detection of proteins binding nonspecifically or nonselectively from those binding specifically to peptides phosphorylated on one individual site or both sites (Figure 1). Identification of Syk Interacting Proteins Mediated by Phosphorylation on Two Tyrosine Residues in Linker B. The analysis of four affinity-purified samples by SDS-PAGE followed by silver staining revealed that the isolated sample mixtures had virtually identical patterns and were so complex that it was difficult to detect specific proteins in individual samples (Figure 2). However, differential analysis was made feasible by the application of iTRAQ-labeling coupled with mass spectrometric analysis. An example is illustrated in Figure 3. Here, the quantitative analysis of two peptides from the same MALDI spot is shown. Peptides of the same sequence labeled with different iTRAQ reagents had identical mass. Therefore, no doublet, triplet, or quadruplet peaks were observed, and no increase in spectral complexity occurred by combining the four samples (Figure 3B). This important characteristic provides a sensitivity enhancement over other isotopic labeling methods that usually result in the observation of peaks in doublet in MS spectra. The top 25 most abundant ions in the single MS mode from each spot were automatically chosen for MS/MS analyses. Illustrated in Figure 3C,D is the selection of two peptide peaks (m/z 1189.7 and 1816.8) for MS/MS experiments. Database search results identified the sequences as YLTVAAVFR Journal of Proteome Research • Vol. 6, No. 1, 2007 135

research articles

Zhou et al.

reduced the interaction between the protein and the phosphopeptide. To verify the specificity of this interaction, we examined the presence of Prdx1 among the proteins that bound to each immobilized peptide by Western blotting using a Prdx1specific antibody. As illustrated in Figure 4, Prdx1 was present in samples purified by both YpY and pYpY affinity resins, but not in samples from the YY or pYY affinity resins. Prdx1 was present in the greatest abundance in the YpY-purified sample, a pattern similar to that revealed by quantitative MS. This result further demonstrates that our approach based on quantitative proteomics is effective in identifying proteins interacting differentially with phosphopeptides containing more than one closely spaced phosphorylation site.

Discussion

Figure 2. Silver-stained SDS-PAGE analysis of proteins isolated from the immobilized peptide affinity purification.

and YSYYDESQGEIYR from tubulin β-2 chain and C2orf4, respectively. Quantitative information was indicated by the ratio of four low-mass signature ions (m/z 114-117) in the MS/ MS spectra. The abundance ratio and Sequest searching identified the C2orf4 protein as a protein that binds preferably to the Syk peptide when the Y346 site is phosphorylated. In contrast, the nearly 1:1:1:1 ratio of 114-117 suggests the abundance of tubulin was not affected by phosphorylation on either tyrosine residue. Therefore, this protein was most likely a background or nonspecifically adsorbed protein. Consistent with the background complexity observed by Silver staining, MS analysis unambiguously identified and quantified a total of 215 unique peptides, representing 76 unique proteins (Table 1 and Supporting Information Table S1). Confidence in these identifications was assigned using open source tools for peptide and protein validation (available from the Seattle Proteome Center Web site: http://www. proteomecenter.org). A peptide probability PeptideProphet score of 0.9 and a protein probability score of 0.8 were chosen as cutoffs. A partial list of proteins is shown in Table 1. The entire list of identified proteins along with quantitative information is available (Supporting Information Table S1). Out of 76 quantified proteins, 65 were present in four samples and showed similar abundances. Of 11 proteins, such as SET, C2orf4, alPha 2, and peroxiredoxin, that showed a change in abundance in binding to four peptides (ratio of 1.5 was used as the cutoff), the majority of them preferred phosphorylated Y346. Only a few proteins bound to the Syk peptide preferably when the Y342 site was phosphorylated (Table 1 and Supporting Information S1). This appeared not due to an improperly immobilized peptide, as Western blotting analyses confirmed the preference of Vav1 for binding to the pYY peptide as reported previously20 (data not shown). However, we cannot rule out the possibility that the immobilization of peptides via their N-termini on the solid phase selectively decreases binding to pY342 as a result of steric effects. Peroxiredoxin 1 Interacts with a Syk-Derived Peptide Phosphorylated on Y346. The quantitative proteomics approach identified peroxiredoxin 1 (Prdx1) as a protein that interacted with the Syk peptide preferably when the Y346 site was phosphorylated (Figure 4B). Phosphorylation on both sites, Y342 and Y346, did not enhance the binding and instead 136

Journal of Proteome Research • Vol. 6, No. 1, 2007

Protein phosphorylation contributes to functional changes in proteins including modulations in their interactions with binding partners. When multiple phosphorylation sites are closely spaced, contributions to interactions can come from individual residues or a combination of residues. We present here a novel strategy based on quantitative proteomics for identifying binding partners for proteins containing multiple phosphorylation sites. This strategy employs a combination of synthetic peptides that are custom-prepared to represent different states of phosphorylation of two closely spaced sites. The use of iTRAQ reagents meets the specific requirements for this strategy by allowing isotopic labeling and analysis of four samples in parallel. Any protein specific for one individual phosphorylation site or a doubly phosphorylated site will be identified by measuring the isotopic ratios in the MS/MS spectrum of the peptide derived from that protein. In addition, the employment of iTRAQ reagents to gain accurate quantitative information in the MS/MS spectra has an advantage over other stable isotopic labeling methods (e.g., ICAT7) which provide quantitative information in the MS spectra. MS spectra are more complicated and have a much higher noise level than MS/MS spectra, making accurate measurements of the relative levels of differentially labeled, low-abundance peptides difficult. For example, the peptide from protein stamin1 tagged with iTRAQ reagents has rather low intensity (indicated with an arrow in Supporting Information Figure S1). If it were labeled with other types of amino-specific reagents, it would have been difficult to obtain quantitative information for this peptide in the MS spectrum. In contrast, quantitative information was exceptionally clear in the MS/MS spectrum with iTRAQ labeling (Supporting Information Figure S1(panel iv)). This approach identified some interesting candidates for Sykinteracting proteins. C2orf4 (Memo/CGI-27/C21orf19-like protein), which bound preferentially to Syk-derived peptides phosphorylated on Y346, was identified previously as a protein that also binds selectively to a phosphopeptide derived from ErbB2.29 This protein, termed Memo in this study, bound to a peptide containing phosphotyrosine 1227, but not to one containing phosphotyrosine 1201. The mode of binding was not determined with certainty, but may have been indirect and mediated by the adaptor protein, Shc. Interestingly, the downregulation of Memo disrupted functional connections between ErbB2 and the microtubule cytoskeleton,29 and Syk also has been reported to bind Shc and is involved in regulation of the microtubule network.30,31 Whether this function requires Syk’s interactions with Memo remains to be explored.

Study of Phosphorylation-Dependent Peptide-Protein Interactions

research articles

Figure 3. Example spectra showing the use of quantitative information to guide identification of specific interacting proteins by µLC and MALDI-TOF/TOF. Components of the spectrum illustrated are (A) total ion chromatogram of peptides recorded on each MALDI spot; (B) MS spectrum of peptides recorded from a single spot (the arrows indicate two ions as the examples that were subjected to MS/MS experiments); (C and D) MS/MS spectra of m/z 1189.7 and 1816.8, respectively (spectra show the dominant fragment ions in the low mass region (m/z 114-117)); (E and F) low-mass regions in spectra C and D were zoomed in to show the signature ions used for quantitation.

Prdx1, which binds preferentially to Y346, also participates in signal transduction pathways by reducing intracellular levels of hydrogen peroxide that are produced by growth factor signaling and by interacting with the transcription factor,

c-Myc, and the protein tyrosine kinase, c-Abl.32-34 Its apparent ability to interact with a phosphotyrosine-containing peptide in a sequence-specific fashion has not been described previously. However, it is known that the active site can accomJournal of Proteome Research • Vol. 6, No. 1, 2007 137

research articles

Zhou et al.

Table 1. A Partial List of Proteins Identified in Immobilized Peptide Affinity Pull-Downs IPI no.

protein

IPI00290770 IPI00290770 IPI00465248 IPI00465248 IPI00465248 IPI00465248 IPI00465248 IPI00465248 IPI00397984 IPI00442522 IPI00442522 IPI00442522 IPI00012507 IPI00012507 IPI00383758 IPI00383758 IPI00383758 IPI00453476 IPI00453476 IPI00453476 IPI00010471 IPI00010471 IPI00010471 IPI00010471 IPI00219018 IPI00219018 IPI00219018 IPI00219018 IPI00219018 IPI00219018 IPI00219018 IPI00219018 IPI00219018 IPI00219018 IPI00219018 IPI00219018 IPI00012048 IPI00012048 IPI00000874 IPI00000874 IPI00000874 IPI00000874 IPI00000874 IPI00003362 IPI00003362 IPI00003362 IPI00025252 IPI00025252 IPI00217223 IPI00217223 IPI00023860 IPI00023860 IPI00023860 IPI00023860 IPI00293276 IPI00293276 IPI00032426 IPI00032426 IPI00032426 IPI00032426 IPI00032426 IPI00032426 IPI00032426 IPI00032426 IPI00026119 IPI00026119 IPI00020984 IPI00020984 IPI00027252 IPI00220740 IPI00012007 IPI00479997

TCP-1-gamma TCP-1-gamma Enolase 1 Enolase 1 Enolase 1 Enolase 1 Enolase 1 Enolase 1 Splice Isoform of SET protein L-lactate dehydrogenase A chain L-lactate dehydrogenase A chain L-lactate dehydrogenase A chain GTP-binding nuclear protein RAN GTP-binding nuclear protein RAN Glyceraldehyde-3-phosphate dehydrogenase Glyceraldehyde-3-phosphate dehydrogenase Glyceraldehyde-3-phosphate dehydrogenase Phosphoglycerate mutase 1 Phosphoglycerate mutase 1 Phosphoglycerate mutase 1 L-plastin L-plastin L-plastin L-plastin Glyceraldehyde 3-phosphate dehydrogenase Glyceraldehyde 3-phosphate dehydrogenase Glyceraldehyde 3-phosphate dehydrogenase Glyceraldehyde 3-phosphate dehydrogenase Glyceraldehyde 3-phosphate dehydrogenase Glyceraldehyde 3-phosphate dehydrogenase Glyceraldehyde 3-phosphate dehydrogenase Glyceraldehyde 3-phosphate dehydrogenase Glyceraldehyde 3-phosphate dehydrogenase Glyceraldehyde 3-phosphate dehydrogenase Glyceraldehyde 3-phosphate dehydrogenase Glyceraldehyde 3-phosphate dehydrogenase Nucleoside diphosphate kinase A Nucleoside diphosphate kinase A Peroxiredoxin 1 Peroxiredoxin 1 Peroxiredoxin 1 Peroxiredoxin 1 Peroxiredoxin 1 78 kDa glucose-regulated protein precursor 78 kDa glucose-regulated protein precursor 78 kDa glucose-regulated protein precursor Protein disulfide-isomerase A3 precursor Protein disulfide-isomerase A3 precursor Multifunctional protein ADE2 Multifunctional protein ADE2 Nucleosome assembly protein 1-like 1 Nucleosome assembly protein 1-like 1 Nucleosome assembly protein 1-like 1 Nucleosome assembly protein 1-like 1 Macrophage migration inhibitory factor Macrophage migration inhibitory factor Protein C2orf4 Protein C2orf4 Protein C2orf4 Protein C2orf4 Protein C2orf4 Protein C2orf4 Protein C2orf4 Protein C2orf4 Ubiquitin-activating enzyme E1 Ubiquitin-activating enzyme E1 Calnexin precursor Calnexin precursor B-cell receptor-associated protein BAP37 Nucleophosmin Similar to Adenosylhomocysteinase Stathmin 1

138

Journal of Proteome Research • Vol. 6, No. 1, 2007

average ratio (114:115:116:117)

1:0.92:1.08:1.07 1:0.78:0.78:0.96

1:0.85:1.56:1.23 1:0.95:1.39:0.94 1:0.85:0.85:0.96 1:0.80:0.82:0.83 1:0.79:0.83:0.82 1:1.05:1.68:1.01

1:1.69:2.43:1.71

1:1.05:1.10:0.98 1:1.30:1.89:1.56

1:0.79:0.89:0.85 1:0.89:0.92:0.97 1:0.79:1.00:1.05 1:0.95:0.95:0.99

1:2:2:2.20 1:1:1.82:1.82

1:0.81:1.11:0.93 1:1.17:1.17:1.15 1:1.32:1.48:1.33 1:1.12:1.00:1.14 1:1.41:1.42:1.36 1:1.00:0.94:0.93

peptide sequence

TAVETAVLLLR GISDLAQHYLMR YISPDQLADLYK AAVPSGASTGIYEALELR FTASAGIQVVGDDLTVTNPKR IGAEVYHNLK LAQANGWGVMVSHR SGKYDLDFKSPDDPSR IDFYFDENPYFENK GEMMDLQHGSLFLR FIIPNVVK LKGEMMDLQHGSLFLR YVATLGVEVHPLVFHTNR KYVATLGVEVHPLVFHTNR VVDLMAHMASKE LTGMAFR LEKPAKYDDIKK VLIAAHGNSLR FSGWYDADLSPAGHEEAKR NLKPIKPMQFLGDEETVRK FSLVGIGGQDLNEGNR MINLSVPDTIDER GDEEGVPAVVIDMSGLR YPALHKPENQDIDWGALEGETR VIISAPSADAPMFVMGVNHEKYDNSLK VIISAPSADAPMFVMGVNHEK LVINGNPITIFQER VPTANVSVVDLTCR GALQNIIPASTGAAK RVIISAPSADAPMFVMGVNHEK LTGMAFR VIPELNGK LVINGNPITIFQERDPSK LEKPAKYDDIKK WGDAGAEYVVESTGVFTTMEK AENGKLVINGNPITIFQERDPSK DRPFFAGLVK VMLGETNPADSKPGTIR LVQAFQFTDK ATAVMPDGQFK GLFIIDDKGILR KQGGLGPMNIPLVSDPKR TIAQDYGVLKADEGISFR AKFEELNMDLFR IEIESFYEGEDFSETLTR IINEPTAAAIAYGLDKR FLQDYFDGNLKR FVMQEEFSR EIVLADVIDNDSWR IKAEYEGDGIPTVFVAVAGR LDGLVETPTGYIESLPR FYEEVHDLER QLTVQMMQNPQILAALQER YAVLYQPLFDKR LLCGLLAER PMFIVNTNVPR NWQDSSVSYAAGALTVH YSYYDESQGEIYR IFILGPSHHVPLSR MSLQTDEDEHSIEMHLPYTAK NGMNMSFSFLNYAQSSQCR IYGELWK TPLYDLR CALSSVDIYR QFLFRPWDVTK LAGTQPLEVLEAVQR LHFIFR WKPPMIDNPSYQGIWKPR LLLGAGAVAYGVR MTDQEAIQDLWQWR ALDIAENEMPGLMR ASGQAFELILSPR

research articles

Study of Phosphorylation-Dependent Peptide-Protein Interactions Table 1 (Continued) IPI no.

protein

average ratio (114:115:116:117)

peptide sequence

IPI00031517 IPI00291006 IPI00299000 IPI00550363 IPI00016342 IPI00000816 IPI00219622 IPI00374151 IPI00101405 IPI00395627

DNA replication licensing factor MCM6 Malate dehydrogenase Proliferation-associated protein 2G4 Transgelin 2 Ras-related protein Rab-7 14-3-3 protein epsilon Proteasome alPha 2 subunit Peroxiredoxin 3 Farnesyl diphosphate synthase Splice Isoform 1 Of Calcyclin-binding protein

1:1.12:1.34:1.02 1:1.15:1.21:1.16 1:1.02:1.13:1.16 1:1:1.12:1.05 1:1.33:1.00:0.91 1:1.11:1.41:1.08 1:1.67:1.02:0.98 1:0.94:1.46:1.34 1:1.12:1.89:1.45 1:0.76:1.59:1.51

DFYVAFQDLPTR IFGVTTLDIVR FTVLLMPNGPMR QMEQISQFLQAAER EAINVEQAFQTIAR YLAEFATGNDRK KLAQQYYLVYQEPIPTAQLVQR DYGVLLEGSGLALR QDFVQHFSQIVR IYITLTGVHQVPTENVQVHFTER

modate a cysteine-sulfinate or cysteine-sulfonate, suggesting that a phosphate group may also be accommodated. This possibility is currently under investigation. While we revealed interacting proteins that are sensitive to individual modification sites, several proteins that have been known to bind to specific phosphorylated residues on the Sykderived phosphopeptides were not identified using this approach.19 It is likely that some binding proteins are present in the cell lysate in amounts that can be detected by more sensitive methods such as Western blotting, but not by mass spectrometry where they might be easily shadowed by highly abundant proteins present in the matrix. We demonstrate through this work that the use of quantitative proteomics combined with synthetic peptides with differential modification sites is a powerful method for finding protein interaction partners and for dissecting the role of multiple sites of covalent modification on protein-protein interactions. This method is particularly useful for identifying

the role of individual sites of modification among multiple closely spaced sites. Quantitative proteomics allows the identification of a small number of specific proteins from the thousands of proteins present in whole cell extracts. This is especially useful under conditions where only mild washing steps are tolerated in order to preserve interactions leaving high background levels of proteins. Specific and nonspecific interacting proteins can then be distinguished with differential isotopic labeling.

Conclusions Quantitative proteomics has seen broad applications in biological research in recent years. The novel strategy presented here is the first report of quantitative proteomics in the identification of the contribution of individual residues to protein-protein interactions. The strategy takes advantage of unique amino-specific stable isotopic labeling reagents known as iTRAQ reagents. Besides several known advantages over

Figure 4. Identification of peroxiredoxin 1 binding to Syk peptide preferably when Y346 is phosphorylated. (A) MS/MS spectrum of peptide LVQAFQRFDK from peroxiredoxin 1. Its characteristic peptide bond fragment ions, type b and type y ions, are labeled; (B) low-mass region (inset) shows the signature ions of m/z 114-117; (C) Western blotting analyses of proteins purified by the immobilized peptide affinity pull-downs using anti-peroxiredoxin 1 antibody, indicating similar quantitative pattern as observed by MS/MS in panel B. Journal of Proteome Research • Vol. 6, No. 1, 2007 139

research articles other amino-specific reagents developed for quantitative proteomics,35-37 in this study, the iTRAQ reagents with four isobaric tags were employed to investigate the role of multiple sites of modification in protein-protein interactions. This feature makes multiplex labeling reagents, such as iTRAQ reagents, most convenient and almost irreplaceable for similar types of analyses.

Acknowledgment. This project has been funded in part by Purdue University and National Institutes of Health Grant CA37372 awarded by the National Cancer Institute (R.L.G.). We acknowledge the use of software in the Institute for Systems Biology developed using Federal funds from the National Heart, Lung, and Blood Institute, National Institutes of Health, under contract no. N01-HV-28179. Supporting Information Available: Figures showing the MS spectra of one MALDI spot and of m/z 1530.8; table showing the complete list of proteins identified in immobilized paptide affinity pull-downs and quantification information. This material is available free of charge via the Internet at http://pubs.acs.org. References (1) Pawson, T. Specificity in signal transduction: from phosphotyrosine-SH2 domain interactions to complex cellular systems. Cell 2004, 116, 191-203. (2) Guo, D.; Hazbun, T. R.; Xu, X. J.; Ng, S. L.; Fields, S.; Kuo, M. H. A tethered catalysis, two-hybrid system to identify protein-protein interactions requiring post-translational modifications. Nat. Biotechnol. 2004, 22, 888-892. (3) Shaywitz, A. J.; Dove, S. L.; Greenberg, M. E.; Hochschild, A. Analysis of phosphorylation-dependent protein-protein interactions using a bacterial two-hybrid system. Sci. STKE 2002, 2002, PL11. (4) Moorhead, G.; MacKintosh, C. Affinity methods for phosphorylation-dependent interactions. Methods Mol. Biol. 2004, 261, 469478. (5) Hinsby, A. M.; Olsen, J. V.; Mann, M. Tyrosine phosphoproteomics of fibroblast growth factor signaling: a role for insulin receptor substrate-4. J. Biol. Chem. 2004, 279, 46438-46447. (6) Patterson, S. D.; Aebersold, R. H. Proteomics: the first decade and beyond. Nat. Genet. 2003, 33 (Suppl.), 311-323. (7) Gygi, S. P.; Rist, B.; Gerber, S. A.; Turecek, F.; Gelb, M. H.; Aebersold, R. Quantitative analysis of complex protein mixtures using isotope-coded affinity tags. Nat. Biotechnol. 1999, 17, 994999. (8) Ong, S. E.; Mann, M. Mass spectrometry-based proteomics turns quantitative. Nat. Chem. Biol. 2005, 1, 252-262. (9) Tao, W. A.; Aebersold, R. Advances in quantitative proteomics via stable isotope tagging and mass spectrometry. Curr. Opin. Biotechnol. 2003, 14, 110-118. (10) Ranish, J. A.; Yi, E. C.; Leslie, D. M.; Purvine, S. O.; Goodlett, D. R.; Eng, J.; Aebersold, R. The study of macromolecular complexes by quantitative proteomics. Nat. Genet. 2003, 33, 349-355. (11) Blagoev, B.; Kratchmarova, I.; Ong, S. E.; Nielsen, M.; Foster, L. J.; Mann, M. A proteomics strategy to elucidate functional protein-protein interactions applied to EGF signaling. Nat. Biotechnol. 2003, 21, 315-318. (12) Ross, P. L.; Huang, Y. N.; Marchese, J. N.; Williamson, B.; Parker, K.; Hattan, S.; Khainovski, N.; Pillai, S.; Dey, S.; Daniels, S.; et al. Multiplexed protein quantitation in Saccharomyces cerevisiae using amine-reactive isobaric tagging reagents. Mol. Cell. Proteomics 2004, 3, 1154-1169. (13) Pawson, T.; Scott, J. D. Signaling through scaffold, anchoring, and adaptor proteins. Science 1997, 278, 2075-2080. (14) Turner, M.; Schweighoffer, E.; Colucci, F.; Di Santo, J. P.; Tybulewicz, V. L. Tyrosine kinase SYK: essential functions for immunoreceptor signalling. Immunol. Today 2000, 21, 148-154. (15) Bolen, J. B. Protein tyrosine kinases in the initiation of antigen receptor signaling. Curr. Opin. Immunol. 1995, 7, 306-311. (16) Turner, M.; Mee, P. J.; Costello, P. S.; Williams, O.; Price, A. A.; Duddy, L. P.; Furlong, M. T.; Geahlen, R. L.; Tybulewicz, V. L. Perinatal lethality and blocked B-cell development in mice lacking the tyrosine kinase Syk. Nature 1995, 378, 298-302.

140

Journal of Proteome Research • Vol. 6, No. 1, 2007

Zhou et al. (17) DeFranco, A. L. The complexity of signaling pathways activated by the BCR. Curr. Opin. Immunol. 1997, 9, 296-308. (18) Kurosaki, T. Genetic analysis of B cell antigen receptor signaling. Annu. Rev. Immunol. 1999, 17, 555-592. (19) Groesch, T. D.; Zhou, F.; Mattila, S.; Geahlen, R. L.; Post, C. B. Structural basis for the requirement of two phosphotyrosine residues in signaling mediated by Syk tyrosine kinase. J. Mol. Biol. 2006, 356, 1222-1236. (20) Moon, K. D.; Post, C. B.; Durden, D. L.; Zhou, Q.; De, P.; Harrison, M. L.; Geahlen, R. L. (2005) Molecular basis for a direct interaction between the Syk protein-tyrosine kinase and phosphoinositide 3-kinase. J. Biol. Chem. 2006, 280, 1543-1551. (21) Simon, M.; Vanes, L.; Geahlen, R. L.; Tybulewicz, V. L. Distinct roles for the linker region tyrosines of Syk in FcepsilonRI signaling in primary mast cells. J. Biol. Chem. 2005, 280, 4510-4517. (22) Hong, J. J.; Yankee, T. M.; Harrison, M. L.; Geahlen, R. L. Regulation of signaling in B cells through the phosphorylation of Syk on linker region tyrosines. A mechanism for negative signaling by the Lyn tyrosine kinase. J. Biol. Chem. 2002, 277, 31703-31714. (23) Law, C. L.; Chandran, K. A.; Sidorenko, S. P.; Clark, E. A. Phospholipase C-gamma1 interacts with conserved phosphotyrosyl residues in the linker region of Syk and is a substrate for Syk. Mol. Cell. Biol. 1996, 16, 1305-1315. (24) Deckert, M.; Tartare-Deckert, S.; Couture, C.; Mustelin, T.; Altman, A. Functional and physical interactions of Syk family kinases with the Vav proto-oncogene product. Immunity 1996, 5, 591-604. (25) Zhang, J.; Berenstein, E.; Siraganian, R. P. Phosphorylation of Tyr342 in the linker region of Syk is critical for Fc epsilon RI signaling in mast cells. Mol. Cell. Biol. 2002, 22, 8144-8154. (26) Eng, J. K.; McCormack, A. L.; Yate, J. R. An approach to correlate tandem mass spectral data of peptides with amino acid sequences in a protein database. J. Am. Soc. Mass Spectrom. 1994, 5, 976-989. (27) Nesvizhskii, A. I.; Keller, A.; Kolker, E.; Aebersold, R. A statistical model for identifying proteins by tandem mass spectrometry. Anal. Chem. 2003, 75, 4646-4658. (28) Keller, A.; Nesvizhskii, A. I.; Kolker, E.; Aebersold, R. Empirical statistical model to estimate the accuracy of peptide identifications made by MS/MS and database search. Anal. Chem. 2002, 74, 5383-5392. (29) Marone, R.; Hess, D.; Dankort, D.; Muller, W. J.; Hynes, N. E.; Badache, A. Memo mediates ErbB2-driven cell motility. Nat. Cell Biol. 2004, 6, 515-522. (30) Nagai, K.; Takata, M.; Yamamura, H.; Kurosaki, T. Tyrosine phosphorylation of Shc is mediated through Lyn and Syk in B cell receptor signaling. J. Biol. Chem. 1995, 270, 6824-6829. (31) Zyss, D.; Montcourrier, P.; Vidal, B.; Anguille, C.; Merezegue, F.; Sahuquet, A.; Mangeat, P. H.; Coopman, P. J. The Syk tyrosine kinase localizes to the centrosomes and negatively affects mitotic progression. Cancer Res. 2005, 65, 10872-10880. (32) Kang, S. W.; Chae, H. Z.; Seo, M. S.; Kim, K.; Baines, I. C.; Rhee, S. G. Mammalian peroxiredoxin isoforms can reduce hydrogen peroxide generated in response to growth factors and tumor necrosis factor-alpha. J. Biol. Chem. 1998, 273, 6297-6302. (33) Mu, Z. M.; Yin, X. Y.; Prochownik, E. V. Pag, a putative tumor suppressor, interacts with the Myc Box II domain of c-Myc and selectively alters its biological function and target gene expression. J. Biol. Chem. 2002, 277, 43175-43184. (34) Wen, S. T.; Van Etten, R. A. The PAG gene product, a stressinduced protein with antioxidant properties, is an Abl SH3binding protein and a physiological inhibitor of c-Abl tyrosine kinase activity. Genes Dev. 1997, 11, 2456-2467. (35) Fu, Q.; Li, L. De novo sequencing of neuropeptides using reductive isotopic methylation and investigation of ESI QTOF MS/ MS fragmentation pattern of neuropeptides with N-terminal dimethylation. Anal. Chem. 2005, 77, 7783-7795. (36) Zappacosta, F.; Annan, R. S. N-terminal isotope tagging strategy for quantitative proteomics: results-driven analysis of protein abundance changes. Anal. Chem. 2004, 76, 6618-6627. (37) Riggs, L.; Seeley, E. H.; Regnier, F. E. Quantification of phosphoproteins with global internal standard technology. J. Chromatogr., B: Anal. Technol. Biomed. Life Sci. 2005, 817, 89-96.

PR0602904